Introducción a Machine Learning para colecciones artísticas
Dentro del programa Patrimonio en un bit, que coorganiza el MUAC (Museo de Arte Contemporáneo de la UNAM) junto con Patrimonio Cultural del Tecnológico de Monterrey, hemos llevado a cabo, junto con el Dr. Martín A. Del Campo, un workshop o taller de introducción a Machine Learning para colecciones artísticas.
El Machine Learning puede ser aplicado a diferentes aproximaciones del patrimonio cultural, en particular revisaremos su aplicación a recursos audiovisuales. El taller comienza con una introducción a los conceptos fundamentales de machine learning o “aprendizaje automático”, y posteriormente se presentarán algunos recursos listos para aplicarse, en los que el asistente podrá analizar sus propias imágenes. Para más detalle del trabajo, les dejo las diapositivas de este workshop de Machine Learning.
Durante el taller, revisamos también las tendencias académicas en Humanidades digitales, encontrando que, verdaderamente, la relación entre machine learning y cultural heritage, encontrando que efectivamente son dos tendencias al alta en los años más recientes:
Ejercicios
Entre los ejercicios realizados, se trabajó primero con la noción de “datos de imagen”, es decir, se observó cómo las computadores codifican las imágenes. Para ello se planteó un ejercicio de exploración de datos de imagen con programación en R. Seguí las pautas que proponen Tilton, L., & Taylor, A. (2015). Humanities data in R. Springer Science+Business Media.
Posteriormente, se planteó un ejercicio más complejo, ya con Redes Neuronales Artificiales para reconocimiento de imagen. También se ejecutó con R, aunque esta vez mediante librearías que utilizan Python “por detrás”. Particularmente, utilicé la librería Keras, siguiendo el ejemplo de la base de datos FMNIST. Se trata de un set de imágenes en baja resolución (28×28) en blanco y negro. Mediante Deep learning el algoritmo clasifica las imágenes según 10 etiquetas (pantalón, deportivas, pullover, camisa, etc.).
En general, el Machine learning es una herramienta poderosa para su aplicación en las Humanidades. Especialmente dentro del patrimonio cultural, hay muchas tareas de reconocimiento y clasificación de recursos.
Architecting Neural Deep Learning Models for Learning Analytics in a Digital Humanities Laboratory
Last week, I participated, together with Dr. Torres-Huitzil, in the 2021 Machine Learning-Driven Digital Technologies for Educational Innovation Workshop, conduced by Writing lab. There, we presented the paper Architecting Neural Deep Learning Models for Learning Analytics in a Digital Humanities Laboratory. The proposal combines the digital humanities approach with the engineering onebuilding the core of an smart environment which was preiously conceived.
The workshop was very well organized by Writing lab, who programmed in two days all the proposals, as well as specific lectures about machine learning and deep learning. Here you can consult the program. The lectures were about Fundamentals of Machine Learning, and Machine Learning in Development of
Educational Technology, both of them by Prof. Amlan Chakrabarti (University of Calcutta) and Dr. Amit Das (University of Calcutta).
On the other hand, the selected tracks were about:
- Digital Technologies for Educational Innovation
- Machine Learning in Educational Innovation
- Inclusion of Digital Technplogies during COVID-19 pandemic
- Virtual and Augmented reality in education
- Reimaging education and Educational Technologies
- Data Science Driven Educational Practices (where our proposal took place).
So, this was a first step for the construction of Neural Deep Learning Models for Digital Humanities Laboratory, a task that is already present in the discussions of several researchers.
Digital Humanities Academic Trends Review: Presentation at 5th Meeting of Digital Humanists
The field of the Digital Humanities has not a very long history, but it has changed significatively since its beginning. Already in 2011, David Berry talked about 3 waves in this emergent area: a first moment of digitalization, when researchers understood technologies as tools to help the management of humanistic resources; a second moment, based on natively digital products, that conceived the tools in their generative mode; finally, Berry intuits a third moment in which the computational revolution would be the motor of the humanistic studies development (Berry, 2011).
Of course, since 2011 many things have happened, especially talking about knowledge areas affected by technologies and their implementations. For example, around 2015 and next years, Big Data was a very powerful trend, which impacted most of the academic research, including humanities and social sciences. Some years later, around 2017, a turn happened with the consolidation of the term Artificial Intelligence, which became even a more paradigmatic trend (Kersting & Meyer, 2018).
Clearly, all these changes have had to influence the Digital Humanities, from where many questions can be formulated: How the academic studies on Digital Humanities have assimilated the Big Data and Artificial Intelligence revolutions? Are the waves detected by Berry still valid or there are needed new ones? What are, more specifically, the trends and approaches that are dominating the most recent studies on Digital Humanities? Even more, what texts and authors were the most influential ones to carry on these transformations?
Interactive Exploratory Analysis
To answer these questions, I propose a Digital Humanities Interactive Exploratory Analysis, reviewing historic academic production . The analysis was performed over the academic database Scopus, from where there were downloaded a total of 3237 documents replying to the query “digital humanities”. With this intention, the collected database was analyzed with R programming tools, mainly with packages such as bibliometrix (Aria & Cucurullo, 2021), widyr (Robinson, 2020) and tidytext (Robinson & Silge, 2020). Some of the applied statistics were: keyword frequencies and its clusterization, co-occurrence and co-citation network analysis, word correlations and their associations within the documents.
For the purpose of socialize the study, I deployed the findings in a public site that will also be the product presented during the meeting. The site will contain the most significant visualizations in an interactive mode, using plotly (Sievert et al., 2021) and shiny (Chang et al., 2021) packages. Furthermore, the network visualizations will be performed by gephi (Bastian & Ramos Ibañez, 2017) and its interactive plugin Sigma JS (Jacomy, 2017).
References
Aria, M., & Cucurullo, C. (2021). bibliometrix: Comprehensive Science Mapping Analysis (3.1.4) [Computer software]. https://cran.r-project.org/package=bibliometrix
Bastian, M., & Ramos Ibañez, E. (2017). Gephi. Makes graphs handy (0.9.2) [Computer software]. Gephi Consortium. https://gephi.org/
Berry, D. M. (2011). The computational turn: Thinking about the digital humanities. The Computational Turn: Thinking about the Digital Humanities, 12. https://sro.sussex.ac.uk/id/eprint/49813/1/BERRY_2011-THE_COMPUTATIONAL_TURN_THINKING_ABOUT_THE_DIGITAL_HUMANITIES.pdf
Chang, W., Cheng, J., Allaire, J., Sievert, C., Barret, S., Yihui, X., Allen, J., McPherson, J., Dipert, A., & Borges, B. (2021). shiny: Web Application Framework for R (1.6.0) [Computer software]. https://cran.r-project.org/package=shiny
Jacomy, A. (2017). Sigma.js (1.2.1) [Computer software]. http://sigmajs.org/
Kersting, K., & Meyer, U. (2018). From Big Data to Big Artificial Intelligence?: Algorithmic Challenges and Opportunities of Big Data. KI – Künstliche Intelligenz, 32(1), 3–8. https://doi.org/10.1007/s13218-017-0523-7
Robinson, D. (2020). widyr: Widen, Process, then Re-Tidy Data (0.1.3) [Computer software]. https://cran.r-project.org/package=widyr
Robinson, D., & Silge, J. (2020). tidytext: Text Mining using “dplyr”, “ggplot2”, and Other Tidy Tools (0.2.4) [Computer software]. https://cran.r-project.org/package=tidytext
Sievert, C., Parmer, C., Hocking, T., Chamberlain, S., Ram, K., Corvellec, M., & Despouy, P. (2021). plotly: Create Interactive Web Graphics via “plotly.js” (4.9.4.1) [Computer software]. https://cran.r-project.org/package=plotly
Wickham, H. (2021). tidyverse: Easily Install and Load the “Tidyverse” (1.3.1) [Computer software]. https://cran.r-project.org/package=tidyvers
The field of the Digital Humanities has not a very long history, but it has changed significatively since its beginning. Already in 2011, David Berry talked about 3 waves in this emergent area: a first moment of digitalization, when researchers understood technologies as tools to help the management of humanistic resources; a second moment, based on natively digital products, that conceived the tools in their generative mode; finally, Berry intuits a third moment in which the computational revolution would be the motor of the humanistic studies development (Berry, 2011).
Of course, since 2011 many things have happened, especially talking about knowledge areas affected by technologies and their implementations. For example, around 2015 and next years, Big Data was a very powerful trend, which impacted most of the academic research, including humanities and social sciences. Some years later, around 2017, a turn happened with the consolidation of the term Artificial Intelligence, which became even a more paradigmatic trend (Kersting & Meyer, 2018).
Clearly, all these changes have had to influence the Digital Humanities, from where many questions can be formulated: How the academic studies on Digital Humanities have assimilated the Big Data and Artificial Intelligence revolutions? Are the waves detected by Berry still valid or there are needed new ones? What are, more specifically, the trends and approaches that are dominating the most recent studies on Digital Humanities? Even more, what texts and authors were the most influential ones to carry on these transformations?
La mirada distante de las Humanidades digitales
TEC_N121
Artículo publicado en la revista Campus Cultural del Tecnológico de Monterrey. Cómo citar:
Cebral-Loureda, M. (2021, September). La mirada distante de las Humanidades digitales. Revista Campus Cultural, 121, 22–26.
La difusión del miedo en Twitter en la pandemia de COVID-19
Aunque actualmente nos hemos adaptado a ella, la pandemia de COVID- 19 ha comenzado como un evento crítico e imprevisto. Sus consecuencias han trascendido el momento. Desde decisiones de salud pública hasta cambios en la producción económica que afectan el núcleo de la vida social e individual. Su irrupción rápida en el contexto mundial, el desconocimiento general de su evolución y los cambios radicales que ha introducido en la vida cotidiana de millones de personas incrementaron el consumo y la circulación de información en la fase inicial de la pandemia.
Las redes sociales han sido consideradas por el periodismo y la consultoría como la puerta de entrada a noticias emitidas por los medios tradicionales acerca de la pandemia. Sin embargo, los usos de Twitter se extienden más allá de ser una red intermediaria entre prensa y audiencia. En los inicios de la pandemia esta red de microblogging se ha constituido espacio para expresión de opiniones, ideas y sentimientos.
Así, por su rol central en la construcción de conversaciones y tendencias, Twitter es la plataforma social más indicada para estudiar la significación social de la pandemia. El estudio se propone identificar los tópicos de mayor interés y los sentimientos asociados en la conversación pública al inicio de la pandemia de COVID 19 en Twitter en Iberoamérica.
El artículo es un resumen divulgativo respecto al original publicado: Cebral-Loureda, M., & Sued-Palmeiro, G. E. (2021). Los inicios de la pandemia de COVID19 en Twitter. Análisis computacional de la conversación pública en lengua española. Cuadernos.Info, 49, 1–25. https://doi.org/10.7764/cdi.49.27467
The beginnings of the COVID19 pandemic on Twitter. Computational Analysis of Public Conversation in Spanish Language
An article written with Dra. Sued-Palmeiro abour Covid-19 in Twitter using R programming and published on the Q2 Journal Cuadernos.Info.
At the beginning of the COVID19 pandemic, social platforms played a crucial role in the production and access to information. This study aims to identify the topics of most significant interest and their associated feelings during the onset of the pandemic on Spanish-language tuits. In addition, we analyzed the role of Twitter as a social platform involved in the public conversation, both as a means for mass self-communication and for amplifying the voice of a reduced set of high visibility actors. 231,375 tweets were collected in Spain and Latin America over two months. Then, the sample was analyzed with digital methods and techniques through computer programming in R. Frequency and sentiment indicators were measured, and terms were grouped to identify topics and determine users’ interests. The frequency of the main terms is dynamic throughout the period studied, suggesting different perceptions of the pandemic. The main topics refer to conversations around the number of cases, deaths, and infections. Sentiment analysis shows the prevalence of negative feelings. The analyzed sample corresponds to ordinary users’ messages for the great majority, but a part of it has been amplified on a large scale through retweets and bookmarks.
Most frequent words for the collected data:
Cebral-Loureda, M., & Sued-Palmeiro, G. E. (2021). Los inicios de la pandemia de COVID19 en Twitter. Análisis computacional de la conversación pública en lengua española. Cuadernos.Info , (49), 1-25. https://doi.org/10.7764/cdi.49.27467
Contention in Deleuze: Fallacies of Connectivist Reading
I am very happy that a new publication is already available on the Online Journal of Philosophy La Deleuziana. It is an special number dedicated to the treatment of the excess within Deleuze philosophy. My contribution has to do with how Deleuze understands the excess only by its opposite, that is, the contention. Otherwise, most of the times Deleuze is consireded a conectivist thinker, something that is not fair with his work.
There is a certain tendency to understand Deleuze under the label of connectivism, a philosophy of life that affirms constant flow, becoming and escaping, a thought that would give rise to an incessant nomadism and, at times, fanatical, constantly dependent on creativity. Sometimes, this is seen as if Deleuze was victim of its own postulates, prisoner of the same requirements he created to run away of social arrangements which, finally, make his thought weaker. Other times the opposite happens: the need for flight, connectivity and creativity is seen as something assimilated by society, something that would align with the own connectivism and accelerationism that promotes the digital and techno-scientific revolution of the 21st century. Against all these approaches, this article tries to show how they are partial. It will be explained how the Deleuzian conception of flow, flight and even overflow cannot be understood except through their opposites -the containment, the asceticism, the sobriety, the impassive or the block of becoming- which are an essential part of his dynamism.
For citation:
Cebral-Loureda, M. (2020). La Contención en Deleuze: Falacias de la Lectura Conectivista. La Deleuziana -online Journal of Philosophy, 12, 54-67.
Artificial Intelligence in Contemporary Philosophy
As a part of the open seminar Science, Technology and Context at Tecnologico de Monterrey and with the international collaboration of Uncuyo, I present the conference The importance of Artificial Intelligence in the redefinition of (some of) the philosophies of the 21st century.
Since Aristotle defined the human being as one who makes a privileged use of reason and the word. This exclusivity of the species has never been more in question than today, in the face of the imminence of new technologies. Artificial Intelligence calls into question many of the alleged unquestionable evidences that define and limit philosophical rationality and forces us to think about our own thinking from new places.
During the session, we will review some of the contemporary philosophies that assume this change of perspective. Among them: the Churchlands, Ray Brassier, David Roden or Reza Negarestani.
Other posts where I treat Artificial Intelligence and philosophy here; which also treat technology implementation from a philosophical point of view.
Nor hermetic, neither hermeneutics: participation at Hermes Congress
I participate on Hermes Congress which takes place at Lanzarote (Spain) from March 22 to March 26. I present a talk about the relations between philosophy and techonology and its configuration in a media lab. The congress want to bring together researchers around the issue of communication, audiovisual media and analysis. As a part of my communication, I present a media laboratory which is an herm3TIC laboratory. With this neologism, I want to approach the hermetica current but trying to opening it throught technologies.
Hermeneutics is one of the most important currents in the human sciences studies. It is opposed to the hermetic current, with less academic prestige and, supposedly, less scientific rigor. However, you can find a large number of authors and studies that support an approach to humanistic studies in a way that is closer to hermetic. In them, the use of technologies is key as part of the implementation of knowledge and communication.
To show this contrast, I expose the theoretical bases of hermeneutics as well as alternative currents closer to hermetic. Then, I reach the notion of “deep time of technology”. This concept brings the most pragmatic tendencies of the human sciences closer to the approaches of the hermetic tradition.
Please, watch the video and share your thoughts about philosophy and techonology relations in the youtube comments 😀! You can find more about this kind of methodologies in previous posts, or even on the herm3TIC project website.